Linear power supply project

[pqass]

The "ref" supply NEEDS to be bipolar (have 3 output posts), that is, have  +12V, 0V/GND, and -12V outputs.
Measuring DCV between +12V and -12V posts with a multimeter should read +24V.
That GND must be connected to +S/+OUTPUT.

My ref transformer is a DP241-5-28 2x14 .400 mA centertap. That will give me my ground for the TL431s, and the fully isolated supply. At least that is what I think will happen.

Using the center tap as GND, am I correct?

Yes. 
See attached (first power supply) and simulation here.
Falstad doesn't have 7812/7912 regulators so I just labelled the place where they belong; follow your full schematic.  Notice the GND symbol on the transformer CT, between bulk caps, 7812/7912 middle pins,  anode of lower TL431, and center output connection.

The +5VREF is derived from the +12 as you've done with resistor+2*TL431 to GND (not the -12V post).

This is good to know because I used -12v for it. I am still not sure how they work to create 5v. But they sure do.

TL431 work like zeners; they shunt/clamp the voltage to their Vz voltage (or programmed via divider resistors on the adjust pin).   The 1K resistor is important otherwise you'll kill the zener/TL431.  However, shunt regulators are only good for low current loads or for reference use (your case).  See here and here. 

The "bulk" supply can be just +12 (to +20V) between 2 output posts (not bipolar).

One of the supplies must be fully isolated from PE; have NO continuity from any output post to PE or the (wide) neutral prong (on a polarized wall plug).  The latter shouldn't really exist but check anyway.

I have the Transformer a 266L24 2x12 and it can be a 2a or 4a depending on using one or two windings selected by the range switch. I understand I can wire it and use it like a center tap. GND the chassis with the PE not the center winding like the REf. Is that correct?

I think I am just going to wire the Transformers and the bridges and at least know my power is correct.

See attached (second "bulk" power supply).   
You still want to keep both ref and bulk supplies isolated from PE; leave that option for the user on the front panel.
Fortunately, the 266L24 secondaries can be wired in series or parallel.  See attached for how to wire a DPDT switch for series/parallel duty (from this "Series/Parallel Switching" diagram).  The 470R resistor (R1) is to discharge the cap when changing the switch position to the lower voltage.   However, you'll likely need another pole (3PDT) for the switch to short R32+R33 in the CREF generation part of the complete schematic (not shown here). 

Wire the primaries for the DP241-5-28 transformer the same way as shown in the Hammond connection diagram (notice the dots on the windings).  In the DP251-5-28 case, F2 connected to F1 is one mains phase and S2 connected to S1 is the other mains phase.      Don't make a mistake or you'll kill the transformer; wire up the mains fuse.

Before proceeding with the regulator parts, wire-up the ref and bulk supplies and confirm that they are producing:
ref:  +12V, +5V, -12V relative to 0V/GND (black multimeter lead).
bulk: +12 or +24V (depending on switch position) relative to 0V (black multimeter lead).

[Therm Mr.]

@pqass there is a lot of good info there I will need to digest. I am going to need to study and think about it for a while before I reply. I thought had a better understanding of T1 from a few previous posts with Kleinstein. But I went back and re-read them I realized I had mis-understood what he said. I only absorbed 1/2 of the reply.

I do have few unrelated questions though.

Fuses F-1 and F-2 call for 1 amp 125v solder in fuses, which I found some Littlefuse for. But it list's F-3 as a fusible resistor 1 ohm 5% .5 watt. Is there a different name for this? Because I can not find fusible resistors anywhere?

And it lists the main T-1 fuse F-4 as a 1 amp or 2 amp in a fuse holder. Size depending on if its OE9 or OE3 and I don't understand what that means. The Transformer is 48 VA which I calculate as .4 amps at 120v so I assume the 1 amp is the correct fuse. Am I correct?

Thanks

Therm

[Kleinstein]

Fusible resistors may be a replacement for low current fuses. For a new build it should be OK to use a suitable low current slow fuse (what ever the transformer can support). The values will depend on the display part.
Fusible resistors are still available, though not longer that common. PTC type fuses can be an alternative in some cases, though the characteristinc and use  is different.

For F1 and F2 it is OK to used exchangible fuses with holder instead of a direct solderable. For DIY this may be easier to get.


The input side fuse depends on the line voltage setting: with 230 V main a smaller fuse is sufficient and should be used. 1 A sounds OK, as the transformer can have some turn on spike in the current, both from the transformer and charging the filter capacitors. 

[Therm Mr.]

See attached (second "bulk" power supply).   
You still want to keep both ref and bulk supplies isolated from PE; leave that option for the user on the front panel.
Fortunately, the 266L24 secondaries can be wired in series or parallel.  See attached for how to wire a DPDT switch for series/parallel duty (from this "Series/Parallel Switching" diagram).  The 470R resistor (R1) is to discharge the cap when changing the switch position to the lower voltage.   However, you'll likely need another pole (3PDT) for the switch to short R32+R33 in the CREF generation part of the complete schematic (not shown here). 

Wire the primaries for the DP241-5-28 transformer the same way as shown in the Hammond connection diagram (notice the dots on the windings).  In the DP251-5-28 case, F2 connected to F1 is one mains phase and S2 connected to S1 is the other mains phase.      Don't make a mistake or you'll kill the transformer; wire up the mains fuse.

Before proceeding with the regulator parts, wire-up the ref and bulk supplies and confirm that they are producing:
ref:  +12V, +5V, -12V relative to 0V/GND (black multimeter lead).
bulk: +12 or +24V (depending on switch position) relative to 0V (black multimeter lead).

I think I had an epiphany moment after reading this. It took some study and thought and a bunch of re-reading of previous posts of @Kleinstein, @andrewtaylor and the post quoted. But I think I am understanding much better now. Please correct me if I am wrong.

My first realization was the switch that shorts R32 and R33 is part of the high low range switch not the CC/CV switch. I have corrected that on my drawing. I have now used a split A/B DPDT switch in those locations.

Second As an AC contractor I have worked with transformers for many years. Low voltage, high voltage with parallel and series primary and secondary windings. Now the parallel series switching you simulated is very complex and difficult to understand. I but I do understand what is going on. It will take some serious though to wire it but I will be able to accomplish that. The parallel low volt high amp and series high volt low amp I get that part.

What was and is still confusing to me, is how HP accomplished the high and low range with the transformer configuration shown? It is completely counterintuitive to me. I understand they had a custom wound transformer made. Is that transformers winding not balanced and one half is higher amp than the other half?

Is it possible to lower the amperage when switching to low range and using the whole winding rather than just one or half a winding? Is the upper winding so much lower in amperage that it lowers the overall when both are in the circuit? 

Sorry if this is a dumb question but it is what is confusing me.

Also I found a triple pole double throw 125v 15a  switch at DigiKey. I think that should be able to accomplish what I need to do. Is that correct or is my epiphany just indigestion:)?

Also thanks for that TL431 info it can be wired in so many ways pretty versatile little component. And for taking the time to create those simulations.

Thanks

Therm

[Kleinstein]

I don't know the details of the HP transformer switching. It is common to have transformer tap switching with just using more or less windings in series, not no extra parallel connection. This does help with reducing the power loss, but does not allow for a higher current at low voltags (at least not much and not higher current limit used). The standard way (e.g. used in many supplies, even the cheap ones) for tap switching is automatic with a relay, based on the actual output voltage. The question is a bit if 2 , 3 or 4 settings are used.

[Wallace Gasiewicz]

I realize that I am repeating Kleinstein but..... My understanding in the dual winding transformer is that the reason for this is to have less voltage lost in the final pass transistors. The pass transistors do not have to bring down the incoming voltage as much and hence there is less heat generated. I have a PS that does this. They use a relay that switches the transformer secondaries. This is automatic in that when you increase the voltage to a certain point there is a sensing circuit that switches the secondaries "automatically". You could use a switch but if you left in in High position and dialed down the volts, the final transistors could get hot.   
There is a bit of a jump in the voltage when the relay switches, this is not entirely a smooth transition.   
The PS I have seen only use one of the secondaries for low volts and then both in series for higher volts.  Have not seen one that uses secondaries in parallel.

[pqass]

What was and is still confusing to me, is how HP accomplished the high and low range with the transformer configuration shown? It is completely counterintuitive to me. I understand they had a custom wound transformer made. Is that transformers winding not balanced and one half is higher amp than the other half?

Is it possible to lower the amperage when switching to low range and using the whole winding rather than just one or half a winding? Is the upper winding so much lower in amperage that it lowers the overall when both are in the circuit? 

The load decides how much current to draw from the secondary (even beyond the amperage spec).   At least until the wire decides to act like a fuse.

HP protects the transformer by limiting what the maximum CREF voltage will be.  See attached and simulation here.

With SW2 shorting R32+R33 (the high current, low voltage position), CREF is high enough to produce 300mV across R23.
With SW2 open (low current, high voltage position), CREF is only high enough to produce 200mV across R23.

300mV is the voltage drop on shunt R2 if 3A flows through it.
200mV is the voltage drop on shunt R2 if only 2A flows through it.
The CREF voltage is the ceiling that kicks-in constant current mode.

We don't know what the wire gauge of the lower (than the tap) secondary winding vs. upper secondary winding. For all we know, HPs custom transformer has a thicker gauge lower winding allowing one to draw more current while keeping under the max. VA.  I do know from another post, the upper and lower windings aren't equal. They measure different ohms.

In my last simulation, I used a DPDT to double the current/half the voltage via parallel/series connections because that is the manufacturers recommend approach.  Their windings are what you'd expect; same length and gauge.

Also I found a triple pole double throw 125v 15a  switch at DigiKey. I think that should be able to accomplish what I need to do. Is that correct or is my epiphany just indigestion:)?

It's more than enough.

[Therm Mr.]

The load decides how much current to draw from the secondary (even beyond the amperage spec).   At least until the wire decides to act like a fuse.

Occam's razor

HP protects the transformer by limiting what the maximum CREF voltage will be.  See attached and simulation here.

With SW2 shorting R32+R33 (the high current, low voltage position), CREF is high enough to produce 300mV across R23.
With SW2 open (low current, high voltage position), CREF is only high enough to produce 200mV across R23.

300mV is the voltage drop on shunt R2 if 3A flows through it.
200mV is the voltage drop on shunt R2 if only 2A flows through it.
The CREF voltage is the ceiling that kicks-in constant current mode.

Great explanation and simulation.

We don't know what the wire gauge of the lower (than the tap) secondary winding vs. upper secondary winding. For all we know, HPs custom transformer has a thicker gauge lower winding allowing one to draw more current while keeping under the max. VA.  I do know from another post, the upper and lower windings aren't equal. They measure different ohms.

This would explain a lot and make a lot of sense.

In my last simulation, I used a DPDT to double the current/half the voltage via parallel/series connections because that is the manufacturers recommend approach.  Their windings are what you'd expect; same length and gauge.

Understood

It's more than enough.

Thank you. I was hoping to use three of the same style rocker switches (though different in function) for three main switches. I not only want a professional quality functioning device, I would also like it to be professional looking. You know both form and function, built with craftsmanship. Function is the obvious priority, so if I need to use a toggle style for the range switch, I will.

My goal has been for this project to replicate the HP 3610 as closely as possible and for it to be an educational and enjoyable experience. Thanks to the folks here on this site, it has been both so far, no doubt.  If I need to make the previously discussed adjustments I will. But before I do I am going to do a few things.

I will look into seeing if I can have the HP transformer replicated relatively inexpensively. Not a solder in but a wire in type configuration. This will simplify the design and build. I have done some research and have become aware of a company, Heyboer Transformer that does small custom work. My understanding is they do this at very reasonable prices. Reasonable being relative I guess. I will see if I can get a quote from them and see if it is reasonable to me.

I will send them the part of the schematic with the transformer. Then what do I ask for the main to be? 2 x 12v hi-range 3 amp, low-range 2 amp. The reference to be 2 x 14/15v 250mA and the display to be either 8/10v 150mA. I don't want to order the wrong thing.

If the above does not work out I will try to contact Hammond and ask if the 266l24 can have a single winding pushed to three amps. Likely the answer would be no (if they even responded) but it can't hurt to ask.

Or maybe just buy a Hammond 266M24 or 266N24 with which a higher amp single winding would support the more simple SPDT wiring. Since the CREF is controlling the output it would not seem to matter if the series connection were higher than two amps. Is this correct thinking or am I off in LaLa land somewhere?


Thanks

Therm

[Therm Mr.]

I was able to contact the custom transformer manu they are nice people. Their prices seem reasonable and very close to the price of purchasing three separate transformers. So far Just ballpark prices so I could have an idea of price. I did not waste their time if too much $$.

I discussed with him what I am trying to accomplish and what I think? I need. But after re-reading @Kleinstein and @pqass so many times and still not always getting it right. I just want be sure I am giving the manu the correct info. I sent the original and my re-draw schematic for his review. Plus I gave him the the info on volts and amps I desired. Below is the discussion starting with his reply after our intial conversation.


Manu Rep, I now come up with  24VAC CT @ 3A winding =72 VA ,  28 VAC CT @ 300mA =8.4 VA , and 5 VAC @ 100mA =5VA  . Total 85 VA transformer. This look correct  ?    I know you draw the higher current at half  the winding( at center tap  ground ) so  am winding  the whole  thing for higher current .  I  can go a bit more for a buffer , never hurts to over build a bit and also use grain oriented laminations, keeps it cooler running.

My reply, was that I think I need more Voltage on the 5v display section for the 7805 regulator to work. Like maybe 7- 8v. And having the extra buffer sounds good. 

His reply, Doing it on a 1 ¼  E center leg  , footprint will be 2 ½ X 2 7/16  4 screws, vertical  end bells.  Arlyn does hard quotes, i just guess best i can. Ill get back with price. I can tic up the 5 volt a bit right now its at  5.36 vac with exact 120 in.

PS, each turn on the 5 volt = .33 volt.  16 turns = 5.36 , 17 turns = 5.69  etc..

After re-reading again I am thinking the primary needs to be a bit higher amperage 3.3 or 3.4 amp. I am not confident enough in my knowledge to buy this transformer without some help from the folks here on this site. So @Pqass @Kleinstein any input?

Right now I will use the transformers I have already purchased to continue testing. By manually switching between series and parallel with wire nuts. I am scrounging for a cabinet or enclosure to mount them in. So I can just run only the secondary wires out to the bridges for safety reasons.  Then I will use them somewhere down the road in something. If not I will consider it the cost of education.

Thanks in advance

Therm.

[andrewtaylor]

So as a safe side recommendation: a 120VA toroid core as a solid  is what I would use.

120VA toroid should be standard value, at least at the manufacturers I buy from it is.

[Kleinstein]

A LED display may need quite some current from the 5 V rail. AFAIR the ICL7106 has some 7 mA per LED and thus ~ 50 mA per digit and thus some 170 mA worst case.  More modern LEDs may get away with less current though.
So I would look more for 300 mA DC  (500 mA AC) for the 5 V part.

To get 5 V via a 7805 regulator it need way more than 5 V AC. The old standard was 10.5 V AC to regulation to 5 V. One may get away with 9 V  and a little less (e.g. 8 V) when using a low drop out regulator instead of the 7805 and accepting some ripple at low mains voltage.

For the main current it need some 1.6 - 1.7 A AC for 1 A of DC. So the 3.4A would be good for about 2 A DC out.

The regulator part can likely get away with less than +-12 V. The current would be mainly for the fan. The regulator part itself needs little current.

[PGPG]

I am shortly at forum. I have no time to read now 150 posts to know everything what was said.
I will just say something that you probably ignore as possibilities are now much higher then when I was doing my linear supply, but who knows may be you will like the idea.
I have done it (first version) when I was about 15 (later I few times made some improvements). Supply was for me first 'must have'. Then whatever I wanted to do I always found that some measurement equipment would be need to do this so each time decision was: let do it first. This way I made analog C meter and frequency meter. Both together with LC generator I used to measure L. The top 'must have' was for me oscilloscope, but I have made it long time later - finished it when I was 24.
In supply I used 100W trafo from tube radio receiver. I disassembled the entire core and unwrapped the anode winding. I replaced it with my winding in a 6+6+6+6+6+1+1+1+1+1 (Volts) arrangement.
All taps are placed on the front panel (wooden plywood casing) so using two banana plugs I can choose any voltage from 1 to 35V in 1V steps and feed it to the rectifier and then the stabilizer (voltage and current). That way I can avoid dissipating too much power in regulating transistor (2N3035 if remember well - 50 years passed). I am still from time to time using this supply.
I wanted to say only about these 6+6+...+1 solution. May be you will like it.
Having these taps at front allow also to use it as AC supply. If it happens that you will need to get high voltage to test something you can use trafo and power its low voltage coil with any 1V..35V AC to get different voltages at its high voltage coil.

[pqass]

For a no-effort, off-the-shelf solution, use your existing Hammond 266L24 for the main transformer and add this Signal 14A-30-512 30VA transformer with dual isolated secondaries (±12VDC@250mA + 5VDC@1.25A after regulation) for the op amp and meter+fan supplies.   See Digikey here.

If you want just one transformer, it wouldn't take much effort to start with a 50VA (open center) toroidal with dual (2x12VAC) secondaries and add your own bifilar 15VAC and 9VAC windings; like the Hammond 182L12 or the Triad Magnetics VPT24-2080  (only one left at Newark/Farnell).  I don't think the additional windings will draw much power away from the existing windings.  And take care not to accidentally create a shorted turn with the mounting bolt.

I'd leave the ±12VDC for the op amps and use the 5VDC for the meters and (5V@200mA) fan.

[Therm Mr.]

For the main current it need some 1.6 - 1.7 A AC for 1 A of DC. So the 3.4A would be good for about 2 A DC out.

This is the part that I was misunderstanding. Initially I thought you meant 1.6 amps AC input to the primary to output 1 amp output to the secondary.

So if am understanding correctly now? For every 1.6 or 1.7 amp AC out of the secondary I will be able to produce 1 amp DC out of the bridge?

This would mean I need an additional 1.6 or 1.7 amps added to the 3.4 amps for a total of 5.1 amps to output 3amps DC after the bridge?

[Kleinstein]

To get 5 3 A DC out after the filter capacitor one would get an AC current of some 5 A and thus need a corresponding winding. A 3.4 A rated winding would be OK for a 2 A supply.

The extra factor comes from the pulsed current and thus an RMS current higher than the averaged current that contributes to the DC current.

[Therm Mr.]

For a no-effort, off-the-shelf solution, use your existing Hammond 266L24 for the main transformer and add this Signal 14A-30-512 30VA transformer with dual isolated secondaries (±12VDC@250mA + 5VDC@1.25A after regulation) for the op amp and meter+fan supplies.   See Digikey here.

If you want just one transformer, it wouldn't take much effort to start with a 50VA (open center) toroidal with dual (2x12VAC) secondaries and add your own bifilar 15VAC and 9VAC windings; like the Hammond 182L12 or the Triad Magnetics VPT24-2080  (only one left at Newark/Farnell).  I don't think the additional windings will draw much power away from the existing windings.  And take care not to accidentally create a shorted turn with the mounting bolt.

I'd leave the ±12VDC for the op amps and use the 5VDC for the meters and (5V@200mA) fan.

 

Yes I would like to simplify down to a single transformer if possible . But I am not sure I am ready to build a transformer yet. Something that is confusing me is if there is a 1.7 to 1 ratio like noted by Kleinstein, how can the 266L24 in parallel @4 amps produce 3 amps after the bridge. There must be something I am not understanding.

I have rounded up an enclosure for my transformers and hope to resume testing with the correct isolated power supplies. Per your suggestions. Regarding the protective earth ground should it land on the transformers metal case? If not where should I land it?

Thanks

Therm

[Therm Mr.]

To get 5 A DC out after the filter capacitor one would get an AC current of some 5 A and thus need a corresponding winding. A 3.4 A rated winding would be OK for a 2 A supply.

The extra factor comes from the pulsed current and thus an RMS current higher than the averaged current that contributes to the DC current.

Did you mean the part highlighted in red to be 3 not five?

Thanks

Therm

[Kleinstein]

Of cause it should be 3 A DC.   ( A add a note to the old post).